US brain cancer rates while cell phones spread

In the eight US cancer registries with records back to 1975, the age-adjusted rate of new brain and other nervous system cancers was 5.93 per 100,000 people in 1975 and 5.66 in 2023, with a high of 7.17 in 1990. Over the same stretch, US wireless connections went from about 340,000 to more than 500 million, and the share of adults who own a cellphone passed 95%. That pattern rules out a large, fast-acting effect of phones on brain cancer across the population. It says much less about small risks and long delays, and a line that lumps all brain cancers together can't speak for any one tumor type.

Where the brain cancer numbers come from

The figures come from SEER, the National Cancer Institute's Surveillance, Epidemiology, and End Results program. Its longest-running group, called SEER 8, covers Connecticut, Atlanta, San Francisco-Oakland, Hawaii, Iowa, New Mexico, Seattle-Puget Sound and Utah, with cases from 1975 onward (SEER registry groupings). A larger group, SEER 12, starts in 1992.

The current numbers are from the release of April 22, 2026, which uses the November 2025 data submission and runs through diagnoses made in 2023 (SEER*Explorer revision history). Rates are age-adjusted, so an aging population doesn't push them up on its own. SEER prints the 2020 rate but leaves it out of its trend model, citing the effect of COVID on that year's data.

Two details matter for reading the line. First, the category counts cancers, which means malignant tumors. Benign tumors such as most meningiomas weren't even reportable to SEER before 2004, according to a SEER registrar ruling, so a series starting in 1975 can't include them. Second, "brain and other nervous system" is broader than the brain: SEER's site definition also takes in the cranial nerves and other nervous tissue.

The broader picture on the same SEER page: about 6.1 new cases per 100,000 people a year for 2019 to 2023, a median age at diagnosis of 61, and a modeled decline of about 1.4% a year from 2014 to 2023. The American Cancer Society projects 24,740 new cases in the US in 2026.

How fast cell phones spread

Two sources track adoption, and they measure different things. CTIA, the wireless industry association, has surveyed carriers since January 1985 and reports estimated subscriber connections. Pew Research Center asks a sample of adults whether they own a cellphone, starting in 2002.

By 2024 CTIA counted 579 million connections, or 1.7 per American, according to its 2025 survey highlights. By then the total was well past the number of people. Pew's survey is the better guide to how many adults own a phone: 62% in October 2002, 91% by November 2015 and 98% in the survey that closed in June 2026 (Pew mobile fact sheet). Pew switched from phone interviews to web and mail in 2023, so the latest points aren't a perfect match for the early ones.

What a flat line can rule out

The simplest reading of the table is that if phones caused a big share of brain cancers, rates should have risen by now. Researchers have tested that idea with numbers instead of eyeballing it.

One detailed US test came from Little and colleagues at the National Cancer Institute, published in the BMJ in 2012 (full text). They took 24,813 glioma cases in non-Hispanic white adults from 12 SEER registries for 1992 to 2008. Over that period, which the authors describe as phone use rising from close to 0% to almost 100% of the population, glioma rates changed by -0.02% a year (95% confidence interval -0.28% to 0.25%). They then projected what the rates would have been if the risks reported in two case-control studies were real:

• With a relative risk as low as 1.5 and a 10-year delay, observed rates should have been higher than they were.
• Using the risks reported in a 2011 Swedish study, predicted rates for 2008 were at least 40% above the observed rates.
• The smaller excess seen only among the heaviest users in the international Interphone study could still fit the US data.

The National Cancer Institute's fact sheet summarizes this line of work: simulations showed that many risk increases reported in case-control studies "were not consistent with incidence data." The 2024 WHO-commissioned review reached a similar point, noting that three simulation studies agreed that relative risks above 1.5 with a delay of 10 or more years were implausible.

What it cannot rule out

The same fact sheet is blunt about the limit. Because trend studies look at whole populations, "their ability to observe potential small differences in risk among heavy users or susceptible populations is limited." Here is what a flat line leaves open:

1. Long delays. US connections passed 100 million around 2000, so registries now cover a little over two decades of mass use. A cancer that took 30 years to appear would not show yet.
2. Small risks. Worked example: a 10% rise on a rate of about 6 per 100,000 adds 0.6 cases per 100,000. In the SEER 8 series the rate has jumped by more than that from one year to the next on its own, from 5.95 in 1984 to 6.96 in 1985, so a rise that size would take years to stand out.
3. Heavy users. If any risk were confined to a small group of very heavy callers, their cases would be a sliver of the total.
4. Tumor location and subtype. Inskip and colleagues, using SEER data for white Americans from 1992 to 2006, found flat or falling trends in every age group except 20 to 29 (full text). The rise among women in their twenties came from frontal lobe tumors, with no increase in the temporal or parietal lobes or the cerebellum, the parts of the brain a phone at the ear would expose most. A separate 2012 study of the Los Angeles County, statewide California and SEER registries found frontal and temporal lobe glioblastomas rising 1.3% to 3.0% a year from 1992 to 2006 while gliomas overall fell; its authors said the cause was unknown and could include changes in diagnosis.
5. Benign tumors. Acoustic neuromas and most meningiomas are outside this series. The NCI cites separate studies showing stable rates for both among US adults since 2009.

Glioma versus all brain cancers

Gliomas are the tumors most often studied in the phone debate, and IARC's 2011 press release put them at about two-thirds of brain cancers worldwide. The SEER category in the chart lumps gliomas together with every other malignant tumor of the brain and nervous system. A rise in one subtype can hide behind a fall in another, which is what the 2012 glioblastoma study found. That's why careful trend tests, like Little's, restrict themselves to glioma and look at age groups and brain regions separately.

What people do with these numbers

Some readers find the registry line reassuring enough. Others prefer to cut their exposure anyway, and the NCI fact sheet passes on the FDA's suggestions: spend less time on calls, use speaker mode or earbuds, and avoid calling when the signal is weak, since the phone then raises its transmit power. Some buy phone stickers sold as EMF protection; EMFCALM, for one, sells them in 3, 6 and 12 packs. Registry data can't tell you anything about a product, and this site doesn't evaluate them.

For now the SEER 8 line ends at 5.66 per 100,000 in 2023, the lowest point in its 49 years. That number describes a population. A question about your own headaches or scans needs a doctor who knows your case.